Monday, July 06, 2009

Open Access: Rethinking Harvard

Last week the architect of Harvard’s Open Access (OA) policy, Stuart Shieber stated: “the Harvard open-access policy could not be, should not be, and is not a mandate.” What are the implications of this for the OA movement in the US?

When in February 2008 it was announced that Harvard University’s Faculty of Arts and Sciences (FAS) had voted unanimously for a resolution to introduce an OA policy Library Journal called it “a shot heard ‘round the academic world.”

Viewed by nearly everyone as the resolution that saw the first Open Access mandate introduced in a US university, the FAS proposal called for all researchers to provide Harvard with “permission to make available his or her scholarly articles and to exercise the copyright in those articles”, and to “provide an electronic copy of the final version of the article at no charge [in order that] the Provost’s Office may make the article available to the public in an open-access repository.”

Since its aim was to ensure that all faculty papers were made freely available on the Internet, the policy was widely hailed as an “historic measure”.  Importantly, it seemed to be exactly the kind of policy that the OA movement had long been calling on universities to introduce.

And while it was not the first such university-level policy in the world, OA advocate Peter Suber pointed out, it was nevertheless “the first in the US, the first to be adopted by faculty rather than administrators, the first adopted policy to focus on permissions rather than deposits, and the first to catch the worldwide attention of the press and blogosphere.”

Certainly the fact that a faculty in a university as prestigious as Harvard had taken such a step was viewed as highly significant, and the news was immediately held to be a major breakthrough for the OA movement. It was also assumed that the policy would be widely emulated by others.

And indeed FAS’ OA policy has proved highly influential. Not only have three further such resolutions emerged from Harvard itself (and a fifth appears to be in the works), but similar resolutions have been successfully passed at Stanford School of Education, at MIT, and at a growing number of other US universities, as well as at universities around the world, including University College in London.

As evidence of its significance, OA advocate Stevan Harnad pointed out to Nature recently that the number of OA policies introduced since the Harvard resolution was passed has “almost doubled globally.” 

Not what it appears?

But what has generally been glossed over, or entirely ignored, is the fact that the Harvard policy is non-binding: So far as giving Harvard permission to distribute their papers is concerned, for instance, any FAS researcher can request a waiver, and be confident that the request will be granted automatically. Moreover, there is no requirement on faculty to deposit papers in the repository, although they are encouraged to do so.

In short, the Harvard policy is not what it appears, or at least it is not what it is represented as. Despite frequent claims to the contrary, for instance, Harvard does not have an OA mandate. After all, if any FAS researcher can obtain an automatic waiver, and there is no requirement to deposit, there is no sense in which the policy could be described as compulsory, and thus mandatory.

This truth was conceded last week by the architect of the Harvard policy, Stuart Shieber. As he put it on his blog The Occasional Pamphlet, “the Harvard open-access policy could not be, should not be, and is not a mandate. I’ve tried to be very careful never to refer to it as a mandate (though I can’t promise I’ve never slipped up).”

Others have surely slipped up: A  press release put out by Harvard last May announcing that the Law Faculty had subsequently also voted for an OA policy, for instance, described it unequivocally as a mandate.

Why is the Harvard policy not mandatory? Because, explains Shieber on his blog, when push comes to shove no one can compel faculty to do anything they don’t want to. “As any dean will tell you, there is no such thing as a mandate on faculty. One could stipulate a policy that all faculty must wear crimson at monthly faculty meetings; the only result would be benign neglect of the requirement by most faculty and assiduous wearing of blue by a small group interested in tweaking the administration. Trust me.”

Does it matter whether an OA policy is technically a mandate? Harnad suggests not: The word mandate, he says, “means both to ‘legislate’ and to ‘legitimize’.”

Presumably Harnad would therefore argue that the Harvard policy is indeed a mandate.

Even if it does include a waiver, Harnad suggests, by legitimising self-archiving the Harvard policy will encourage faculty to do something that until now they have assiduously chosen not to do.

In other words, it doesn’t matter whether it’s called an OA policy or a mandate – so long as it persuades researchers to make their work freely available.


But is there not a danger of fudging the issue here? While I make no claims to be a lexicographer, no definition of the word mandate I can find suggests that it also means to legitimise. Rather, the consensus seems to be that, in the words of the Merriam–Webster dictionary, a mandate is, “an authoritative command; especially: a formal order from a superior court or official to an inferior one.”

And if one goes back to the Latin root of the word (courtesy of Perseus’ digital version of the canonical Latin Dictionary at Tufts University), one finds that the verb mando means “to commit to one's charge, to enjoin, commission, order, command.”

Of course one can always argue about the meaning of words, and I don’t meant to be overly pedantic. But this isn’t only a question of semantics. In calling for mandates OA advocates have always assumed (or certainly implied) that self-archiving would need to be made compulsory, not simply legitimised.

And they did so for a good reason: mandates are necessary, they rightly argued, because researchers won’t self-archive voluntarily, however legitimate they believe it might be to do so.

As Harnad frequently puts it, “surveys show that only 15% of authors will self-archive unless it is mandated. Just requesting or recommending deposit does not work. Deposit analyses comparing mandated and unmandated self-archiving rates have shown that mandates (and only mandates) work, with self-archiving approaching 100% of annual institutional research output within a few years. Without a mandate, IR content just hovers for years at the spontaneous 15% self-archiving rate.”

Carrots and sticks

This is not just supposition, as Harnad points out. There is hard evidence to support the claim that non-mandatory policies are invariably unsuccessful. We know, for instance, that for so long as the NIH Public Access Policy remained a request compliance levels were derisory.

When the policy was upgraded to a requirement, however, deposits immediately began to grow rapidly. As an NIH spokesperson put it to me recently, since the policy became mandatory, “Compliance has increased almost 250% ... It has jumped from 19% of our target estimate 80,000 papers per year arising from NIH funds during the voluntary policy to almost half (49%) of the target estimate of papers arising from NIH funds at the end of 2008.”

That said, while Harvard’s policy is not a mandate it has at least shifted the emphasis from opt-in to opt-out. The question is, will this be sufficient to ensure greater compliance than a voluntary policy? If so, how much greater?

Of course, the proof of the pudding will be in the eating, so perhaps we shouldn’t prejudge the matter. For now we don’t know what level of compliance the Harvard policy might achieve. And since – nearly eighteen months after Harvard’s OA policy was agreed – its repository has yet to be made publicly available we have no idea how many papers are currently being deposited.

So far as Shieber is concerned, however, Harvard’s policy is the best that can be hoped for at the university level. “I am not claiming that there can be no true open-access mandates on faculty,” he says. “Rather, such mandates must come from outside academia. Funders and governments can mandate open access because they can, in the end, refuse to fund noncompliers. They have a stick. All a university, school, or dean has, in the end, is a carrot.”

The trouble is that if it is not possible to impose mandates at university level in the US the OA movement faces a worrying obstacle. It is also confronted with an awkward question: As Harvard-style policies propagate across North America might the movement discover that its boat has developed a leak under the waterline?

All in all, for the moment OA advocates in the US might be better to focus their energies on persuading research funders and the US government to impose mandates, and not allow themselves to be distracted by the string of announcements coming out of universities, some of which are undoubtedly more hot air than substance.

For this reason initiatives like the recently re-introduced Federal Research Public Access Act (FRPAA) are doubtless of much greater potential significance.

If successful, the FRPAA would require all US agencies with research budgets of $100 million or more to ensure that the output of any research they fund is made freely available within six months of publication. And we can be confident that the FRPAA will arm these agencies with a stick, not a carrot.

Certainly anyone who believes that university-level OA policies are on the verge of sparking a revolution in the US might need to rethink the matter. And in the process, they may have to conclude that the current poster child of the OA movement is not all it claims to be.

On the other hand, Harvard might yet surprise us all!


Stevan Harnad said...

OED on "Mandate, n.":

I. A commission
II. A command; an authorization

That said, I agree with Richard that it would do OA no service to water down Green self-archiving mandates from requirements to requests. Mandates need to legitimize as well as legislate self-archiving.

DAW said...

An example of an OA Mandate (though departmental and not university-wide) that DOES have teeth is that of the University of Oregon Department of Romance Languages, which requires that URLs of postprints be included in materials submitted for review and promotion:

Richard Poynder said...

Thanks for this David.

So it is possible to impose a self-archiving mandate on a US faculty?

Are you able to indicate what level of compliance the RL mandate is achieving, or is it still too early?

Stevan Harnad said...


Several of the universities that have mandated deposit have linked their policy to performance review, e.g., University of Liège:

   "as announced a year ago in March 2007, starting October 1st, 2009, only those references introduced in ORBi [U. Liège's Institutional Repository] will be taken into consideration as the official list of publications accompanying any curriculum vitæ in all evaluation procedures 'in house' (designations, promotions, grant applications, etc.)." ( B. Rentier, Rector, U. Liège)


Exerpt from What's in a Word? To "Legislate" and/or to "Legitimize": the Double-Meaning of (Open Access) "Mandate":

No penalties for noncompliance. All experience to date has also shown that whereas adding various positive incentives (rewards for first depositors, "cream of science" showcasing, librarian assistance and proxy-depositing) to a mandate can help accelerate compliance, no penalties for noncompliance are needed...

"Publish-or-perish." sanction/incentive (depending on how you look at it) that is already implicitly built into the academic reward system: Is "publish-or-perish" a mandate, or merely an admonition?

Research impact. Academics are not "required" to publish, but they are well-advised to do so, for success in getting a job, a grant, or a promotion. Nor are publications merely counted any more, in performance review, like beans. Their research impact is taken into account too. And it is precisely research impact that OA enhances.

Performance review. So making one's research output OA is already connected causally to the existing "publish-or-perish" reward system of academia, whether or not OA is mandated. An OA mandate simply... makes the causal connection explicit... [A] number of the mandating institutions have procedurally linked their deposit mandates to their performance review system as follows:

Submission format. Faculty just about everywhere already have to submit their refereed publication lists for performance review today. Several of the universities that mandate deposit have simply updated their submission procedure such that henceforth the official mode of submission of publications for performance review will be via deposit in the Institutional Repository.

"Online or invisible." This simple, natural procedural update -- not unlike the transition from submitting paper CVs to submitting digital CVs -- is at the same time all the sanction/incentive that academics need: To borrow the title of Steve Lawrence's seminal 2001 Nature paper on the OA impact advantage: "Online or Invisible."

Keystroke mandate. Hence an OA "mandate" is in essence just another bureaucratic requirement to do a few extra keystrokes per paper, to deposit a digital copy in one's institution's IR. This amounts to no more than a trivial extension to academia's existing "mandate" to do the keystrokes to write and publish the paper in the first place: Publish or Perish, Deposit to Flourish.